A novel K-means classification method with genetic algorithm

Abstract

Data classification is an important part in data mining field. However, problems of high amount of calculation and low accuracy always existing in data classification attract interests of many researchers. This paper proposes a K-Means classification method with genetic algorithm applied to faster and more accurate classification. A data preprocessing approach based on sorted neighborhood method (SNM) is designed to clean the redundancy data effectively. The K-Means method is then utilized to classify the processed records. In order to improve the efficiency and accuracy, the genetic algorithm (GA) is applied into K-Means model to perform the dimension reduction. The results of simulations and experiments demonstrate that the proposed method has better properties in efficiency and accuracy than the competing methods.